Open benmoss opened 1 year ago
For the purposes of allowing a compact representation of multiple test cases from a single vantage point I like this model. While discussing this additional expressiveness and looking at the project linked above @benmoss pointed out that the model it uses permits many degrees of freedom, probably more than we need.
edit: convo about allowed/denied moved to #5
These are separate issues, but it's my mistake for including the allowed
/denied
change here. I've updated the original post to disentangle the whole "protocol semantics" problem from the multiple probes problem
In the current config each test case runs a single probe which tests a single
from->to
relationship. @rsalmond and I were chatting about how we might able to express multipleto
s from a singlefrom
.We thought maybe something like this:
where a
from
has manyto
(replacing the singleto
). There's a lot of ways to shave this cat, we came across this project that basically is trying to achieve the same thing with a very different config style.